Cake Job Search

公司介紹 我們是一家深耕金融科技與資料智能應用的技術團隊，正積極將 大型語言模型導入高風險、高標準的實際業務場景。不同於單純模型實驗，這裡更關注 AI 是否「可靠、可驗證、可長期運作」。 團隊由資深工程與資料背景成員組成，文化務實、重視工程品質，並持續投入 AI Safety、模型評測與自動化流程，打造能被真正信任的生成式 AI 系統。 工作內容 設計並開發 LLM / RAG 系統的自動化評測工具與流程 建立模型多維度評分機制（Accuracy、Relevance、Faithfulness、Consistency） 將評測流程整合至 CI/CD（Jenkins / GitLab CI），確保模型版本品質穩定 與業務及資料團隊合作，建立高品質 Golden Dataset 與 Synthetic Data 執行 Prompt Injection、Jailbreak 等 LLM 安全與紅隊測試 分析失敗案例，協助定位 Retrieval 或 Generation 問題，驗證修復成效 使用的技術 Python、Pytest / unittest LLM / RAG Framework：LangChain、LlamaIndex、Semantic Kernel 資料分析：Pandas、NumPy、Visualization Tools 評測框架：RAGAS、TruLens、DeepEval、Promptfoo（加分） CI/CD、Git Docker / Kubernetes（加分）

As an SDET at JKOpay, you will collaborate with RD and PM teams to conduct test development in your area of responsibility according to the project's progress. In addition to project tasks, you will assist with app-related release matters. Lastly, you will work together to maintain technical projects specific to SDET and optimize existing frameworks and processes.We expect you to possess strong communication skills, teamwork abilities, problem-solving capabilities, resilience under pressure, adaptability, proactivity, and technical expertise. Additionally, you should have a passion for and a solid understanding of the electronic payment market, and be willing to fully dedicate yourself to the product.▶ Programming is required. • Overall quality on both the back-end side and the front-end side with the creation of test plans, cases, and reports. • Create and maintain reliable automation scripts• Ensure quality of App release by understanding new features/functions and verifying various fixes.• Review requirements/design documents and propose recommendations.• Implemented a testing tool to accelerate the testing cycle.